Crack Paths 2009

strength should lead to lower F C Grates. Typical locations in the yield profiles (Figure

2) were chosen for the hole in FS welded AA2024-T3and AA7075-T6(Table 1).

For the 0º configuration no significant change of crack path is expected, because the

residual stresses are parallel to the applied load and do not lead to a rotation of the

principle stress at the crack tip. Nevertheless, a change in F C Grate is expected because

the residual stress adds up to the applied load. Interesting for this configuration is the

comparison between the F C Grates in AA7075-T6and AA6013-T4, because the latter

material has no change in yield strength in the FS weld, but only a residual stress field.

The 45º configuration was expected to show changes in the crack path because the

residual stress should result in a rotation of the principle stress and thus the orientation

of the crack path. Moreover, the yield strength profiles of the FS welds will change the

orientation of the plastic zones and thus influence the crack path orientation.

Table 1, configurations of specimens

Material:

Configuration: Distance hole to weld centre [mm]: Sawcut:

45º FS weld 8.3 (for all)

1-side 2-side

AA2024-T3 90º FS weld 0.0; 3.0; 6.5; 6.5

21.6; 16.6

0º FS weld

1-side

45º FS weld 8.3 (for all) 7 0; 10.0

AA7075-T6 90

1-side 2

19.0; 10.9

AA6013-T4 0º FS weld

1-side

Digital image correlation

To obtain the strain field around the crack tip, (DIC) was used [8,9]. The software to

process the images was developed by the author [10]. The images of 1600 by 1200

pixels were captured at certain fatigue life intervals, using a C C Dcamera which was

positioned in front of the specimen and aligned perpendicular to the specimen’s surface.

To obtain a high accuracy (large amount of pixels/mm2) but also a large measurement

area, a grid of images were taken with a small overlap using an x-y-z controlled rack.

The reference image was taken at a load of 1 kN, while all the subsequent images were

taken at the maximumfatigue load. The measured strains were corrected afterwards for

the applied load in the reference image.

From the deformation measured in the images, the strains in two directions were

obtained, i.e. x- and y-strain. For this research only the y-strain is of interest, because it

is oriented in the direction of the applied load. To obtain the strain in the images a grid

was used with an initial distance between the grid points of 60 pixels (L0). The strains

are calculated for each image individually after which the images and the strain data are

fitted together to obtain the full measurement area.

R E S U L T S

The most interesting observations made during this study are presented in this chapter.

506